The present invention relates to a technique of digital watermark (herein-after “watermark”) for a motion picture.
Watermark technique is drawing more attention from the viewpoint of copyright protection for digital data such as a motion picture. Watermark technique is technique of embedding specific information in digital data according to predetermined rules such that the mentioned specific information can-not be extracted from the digital data at least without using the mentioned predetermined rules. For example, according to predetermined rules, information on a purchaser or the like of image data is embedded in an invisible form into the motion picture itself. When the image data is copied illegally, a person who has copied illegally can be specified by using the predetermined rules to extract the embedded information from the illegally copied data.
As well as still image data, motion picture can be an object of embedding watermark information. In that case, it is considered that watermark information is embedded in field data as a constituent of motion picture.
However, there exist various scanning methods and various picture sizes for motion picture. As a scanning method, there exist an interlace method in which 2 fields are scanned to give 1 frame, or a progressive method in which 1 field is scanned to give 1 frame, or the like. Further, as a frame size, there exist 352×240 dots employed for example by MPEG (Moving Picture Experts Group)-1, 720×480 dots employed for example by MPEG-2, 1920×1080 dots, or the like.
In order that watermark information can be detected correctly for motion picture of all scanning methods and all picture sizes, it is necessary to determine a memory capacity for storing field data used in detection of watermark information, considering a scanning method and picture size for which a field data size becomes biggest. Accordingly, the memory capacity becomes large, and hardware size of an apparatus for detecting watermark information becomes large.
Further, there is a possibility that, in the distribution process, motion picture is variously changed in scanning method and picture size, for example. It is required that watermark information is detected correctly even from motion picture that has been subjected to such changes. In this respect, Japanese Unexamined Patent Laid-Open No. 2001-053955 proposes a technique of correctly detecting watermark information from image data that has been subjected to geometric transformation.
However, when this technique is used to detect watermark information from modified motion picture, detection algorithm becomes more complex as the number of scanning methods and the number of picture sizes become larger, and accordingly, a processing time required for detection of the watermark information becomes longer. For example, when motion picture on the market employ either the interlace method or the progressive method as a scanning method and a picture size of either 720×480 dots or 1920×1080 dots, there exist a plurality of patterns of changes that may be made in the distribution process. The number of patterns depends on the product of the number (2×2=4) of combinations of a scanning method and a picture size employable in embedding watermark information into motion picture and the number (2×2=4) of combinations of a scanning method and a picture size employable by the motion picture at the time of detecting the watermark information. Thus, when it is tried to detect watermark information assuming every change pattern, as the number of scanning methods and the number of picture sizes are larger, the number of change patterns increases at a squared rate and a processing time required for detecting the watermark information becomes longer.